Tuesday 22 October 2013

Every Tech Community Needs To Exclude Sexism, So Constraint Programming Should Exclude Sexism

Update 23 Oct: I am thrilled to say that the very next day after getting my email, the local organisers of CP 2014 (Christine Solnon and Yves Deville) have agreed to implement an anti-harassment policy, and to investigate child care possibilities for the conference. That's my points 1 and 3 immediately agreed to, and I couldn't be more thrilled.  Many thanks. 

It's not too often that syllogisms are useful in everyday life, but this is one occasion.
  • Every tech community must eliminate sexism.
  • Constraint programming is a tech community. 
  • Therefore constraint programming must eliminate sexism.
I've been in constraint programming for almost 20 years, but it's only recently that the necessity of the first point was brought home to me, because of what's been happening in the Ruby community.

Here's a tweet: 

    You know what? Women are seriously worrying if it's ethical to advise women to go into Ruby.  Here's a followup to that tweet, from spinnerin
    "We need allies to speak up and show their support for a Ruby community that is welcoming to all participants, not only the privileged majority."
    If you don't know why this has recently come to the fore in the Ruby community, lucky you.  Put it this way: somebody has deleted her blog post about her sexual assault at codemash, because she does not want her claim to fame to be the woman who was sexually assaulted at codemash

    I'm not in the Ruby community, but I am in the Constraint Programming community. It's a great community, and many of my best friends are in this community, men and women. But that doesn't mean we can't be better.  We can be better and we must be better.

    As a man in the constraint programming, it's time to stop worrying about appearing to be politically correct, and it's time to stop thinking that maybe sexism isn't a major problem in constraints because we are all nice people. It's time to start being the first line of defence for women.  It's time to start figuring out how to make women's time in constraints better, to the benefit of those women (because working on constraints is super-fun) and for everybody (because women do superb work in constraints.) 

    Here are some thoughts on how we can make constraints - and especially the constraint programming conference - a better place for women and other historically disadvantaged groups.  The next conference will be in Lyon, in 2014, and I have some specific suggestions. 

    Before I get to this, let me say that I don't think Constraints is a particularly bad research area for women.  Apart from anything else, the third winner of the ACP research excellence award was a woman (Rina Dechter), and the first winner of the ACP distinguished service award was a woman (Francesca Rossi). On the other hand, it's not that great either. This year, for example, the Programme Committee of CP 2013, counting Senior and Technical committees, comprised 61 people, if I counted right. The number of women in the world on this committee? 4.  The number of men on the committee who work on the same corridor as me in St Andrews?  4.  And this is from a programme chair (Christian Schulte) who ran a brilliant programme and who I would be astonished to find was sexist.

    1. Implement a Conference Anti Harassment Policy

    Honestly, this is a no brainer. We should have a conference anti-harassment policy. This is not just for the protection of women attendees. Conference volunteers need to know what to do if something bad happens.  But let the good people at geekfeminism make the case:

    "Why have an official anti-harassment policy for your conference? First, it is necessary (unfortunately). Harassment at conferences is incredibly common - for example, see this timeline of sexist incidents in geek communities. Second, it sets expectations for behavior at the conference. Simply having an anti-harassment policy can prevent harassment all by itself. Third, it encourages people to attend who have had bad experiences at other conferences. Finally, it gives conference staff instructions on how to handle harassment quickly, with the minimum amount of disruption or bad press for your conference."
    And it's not hard to write an anti-harassment policy, because you can copy and paste it

    We should make CP 2014 the first Artificial Intelligence conference on this page (though perhaps other confs have had a good policy but have not made it to that page yet.) 

    If we can't do this, we can't advise women to attend CP 2014 or any other future conference.  

    2. Make Sure At Least One Woman Is An Invited Speaker 

    At CP 2009, there were two female invited speakers out of three.  They were Barbara Smith and Carla Gomes.   In the four years since there have been thirteen invited talks, with one female speaker (Michela Milano in 2013).  One of 13 is not great is it, people?

    One invited speaker per year in a historically disadvantaged group is not a lot to ask. Yes, I know, this starts to sound like positive discrimination, doesn't it? Maybe you don't like positive discrimination. 

    But think about it this way.  Do we need to work to make sure at least one speaker is a man? No. Every single year in the history of CP, at least one invited speaker has been a man. (Believe it or not, I've checked this statement, even though it hardly needed checking.)  On the other hand, in three of the last four years there have been zero women.   And since 2009 is the only year there was more than one woman, that's the only year in history that women outnumbered men among invited speakers at the major constraint programming conference.  

    Oh, by the way, if you really really can't think of any other women to invite, ask some of the same ones again!  They're great.  Oh I know, we couldn't have the same ones again, could we? Well, not unless they were men like Moshe Vardi (2002 and 2010) or Peter Stuckey (2005 and 2013) or Pascal van Hentenryck (1997 and 2013) or Helmut Simonis (2004 and 2006) or Alain Colmerauer (2001 and 2008) or George Nemhauser (1996 and 2002).  That's twelve invited talks by men who have given two invited talks each. The same number of men have given two invited talks as women who have given any. Compared with six invited talks by women.  

    You really don't want the same ones again?  Still stuck?

    What about Zeynep Kiziltan? She'd be great.  Or Inês Lynce? She'd be great. Or Karen Petrie? She'd be great. None of these are junior people. You might have better ideas since I've only picked people I'm close to (written papers with Karen and 
    Inês, highly connected to Zeynep via Ian Miguel and others.) 

    Update 22 Oct 2013: What I've said here is stronger than what is recommended by this FAQ at "forgenderequityatconferences". I certainly agree with that FAQ that if good faith effort to invite women is made, that would be acceptable even if no woman eventually appears.  However I'm leaving the text above as it originally was because it's what I wrote and I do think that good faith efforts could be rewarded with success, at least for next year.

    Some disclosure: I was chair of CP 2009 and invited Barbara and Carla.  But I claim no credit for inviting women, only for inviting people who I thought would be the best.  Carla had just started up  Computational Sustainability, a very hot topic and was recommended by more than one member of the programme committee. And if I was biased towards Barbara, it was as an old friend who had (incomprehensibly to me) never been asked to give a keynote despite being the best constraint modeller in the world. 

    3. Make sure childcare is available

    Are you prevented from going to a conference because of childcare problems? No? Great! Let's make sure nobody else is either. If you have a child, they might love attending CP. But if you have to travel alone, maybe because you are a single parent, you need to know that childcare is available. This is a gender-blind problem but it is likely to differentially affect women.  

    I don't think CP 2014 needs to have a childcare centre on site. But I do think that we should make sure that proper childcare facilities are available and linked to.  What this means depends on the local organisers, so they might be a childcare centre, or there might be trustworthy carers who could do childcare for a fee.   

    Hey, you know what, I know most people don't need this.  And people could organise it themselves... but ... everybody could organise their own hotel.  It's still nice to have a bunch of conference hotels to make our lives easier.  Same thing with childcare. Though of course finding reliable childcare in a strange city in a foreign country is much harder than finding a decent hotel.  If childcare makes the difference between attending or not, having it linked to from the conference site could make the difference between attending or not, and then attending that inspirational talk or making that critical contact which changes your career.

    4. Make it affordable if you're disadvantaged

    What is one of the great successes of the CP community?  I would say it was the CP Doctoral Program. This has helped literally hundreds of people attend CP, people who were potentially disadvantaged by being junior people studying for a PhD and without their own funds. By reaching out to these people we have brought them into the community to meet other researchers.  The CP community has been happy to help them financially, by accepting larger conference fees for everybody else.

    What a great and wonderful thing! 

    Before you ask, no I don't want lower conference fees for women. 

    But financial help with childcare? Help for disabled people, like maybe paying for travel for a helper? Yes. Let's help them just like we have helped so many graduate students. It won't cost much, and if it did cost a lot it would be fantastic for the community.

    5. Let's celebrate the Petrie Multiplier

    I recently wrote about the Petrie Multiplier. This blog post has had about 25,000 page views (which I understand to be a lot for a noname blogger), and been tweeted about hundreds of times. It's a simple model which explains how sexism is magnified by the gender disparity in a field like tech (or subfield like constraints).  But you know what, it's an idea invented by a constraint programmer (Karen Petrie) and written about by a constraint programmer (me).  

    Let's understand that in constraints, where women are in the minority, just being in the minority can lead to dramatically more sexism per woman, and to men not seeing that sexism. 

    This is not a major contribution from constraints to eliminating sexism, but it is something.  Let's celebrate it, by making sure we are one of the communities that leads tech in getting rid of sexism.  

    6. Make a commitment to gender equity at conferences like Constraint Programming 

    I've signed the commitment to gender equity at conferences. Please do so too. There's an FAQ available.

    7. Let's not argue about whether sexism is a problem

    I'll be really happy to argue about how to fix the sexist problem we have. I don't even mind if you don't call it sexism (call it "gender imbalance" or "inclusiveness" or anything you like).  

    But please don't argue that we don't have a problem. Here's the best two ways I can put this point. 

    I started thinking I should write this post a week or two ago.  Or rather, I thought I should send a letter to the most relevant people (organisers of CP 2014.)  But it's too important for that, so I'm posting it here.

    Update: here is the letter I actually am sending to the relevant people: 

    Dear Yves, Christine, Barry, Pierre and Helmut 

    As local chairs and program chair of CP 2014, and conference coordinator and president of the Association for Constraint Programming.   

    I wish to propose that the Association for Constraint Programming commit to  moving towards gender equality at CP 2014 and all future conferences. 

    By far the most important point (in my opinion) is to adopt a strong anti-harassment policy, to prevent events such as have recently been tearing the Ruby community apart.   But there are other good things we could and should do.

    You can read my reasons for suggesting this, and concrete suggestions on my blog at http://iangent.blogspot.co.uk/2013/10/every-tech-community-needs-to-exclude.html

    Very best wishes



    1. Here is a little story. Once upon a time in the bad dark ages a sexist professor gave a plenary at a conference about a technical device he had chosen to call a "zipper". With bad puns. And gestures. Someone stormed out and folks applauded. The professor later apologised in various public fora. But since he recalled the event as his best moment of the conference in 25 years, and the organisers duly put it in the commemorative slide deck, one wonders exactly how contrite he and the organisers were.... (See slide 25) http://lii.rwth-aachen.de/lics/lics25th.pdf

      So the moral is: state the values and live the values guys....

      1. Thanks Anon.

        That incident is really shocking and is now recorded at geekfeminism


        It's also a great example of why we need harrassment policies. With a harrassment policy in place somebody (man or woman) could have stood up to say it was unacceptable, and known they would be backed up by the conference.